Patients selection
The study was approved by the ethical committee of the IRCCS Rizzoli Orthopedic Institute (protocol number 11551/CE/US/ml, 5 May 2006).
A cohort of 31 patients with knee osteoarthritis (OA) candidate for TKA was enrolled for the present study after signing an informed consent between 2011 and 2012.
The inclusion criteria were: (1) Primary knee osteoarthritis, (2) Kellgren-Lawrence grade 3–4, (3) BMI < 40 kg/m2. The exclusion criteria were: (1) Previous lower limb alignment corrective surgery on the affected side, (2) BMI > 40 kg/m2, (3) Rheumatoid arthritis, (4) Post-traumatic arthritis.
The mean age of the patients included in the study was 70.5 ± 6.5 years (range 83–54 years), 9 males and 22 females.
Acquisition protocol
All the patients underwent intra-operative kinematics assessment with a commercial navigation system (BLU-IGS Orthokey, Lewes, Delaware) equipped with a software specifically focused on kinematic analysis (KLEE, Orthokey, Lewes, Delaware) [14]. This system has a 3D RMS volumetric accuracy of 0.350 mm and a 3D RMS volumetric repeatability of 0.200 mm [21], as reported by the producer. All the kinematic data were off-line processed by applying proprietary routines developed in Matlab (Mathworks, Natick, MA, USA).
The proposed methodology was assessed to have a repeatability lower than 2 mm in translation and lower than 3° in rotations [14], with ICC values ranging from 0.94 to 0.99 [4].
Anatomical landmarks on femur and tibia were acquired to define the joint coordinate reference system (JCS) [5, 8] and to perform TKA navigation protocol. The anatomical registration on the femur consisted of: the femoral head (by leg pivoting), the most distal part of the femur in the intercondylar notch (over to the lateral margin of the posterior cruciate ligament), the anterior shaft, the medial and lateral epicondyles, the most posterior and distal part of the condyles and the Whiteside Line (WSL). The medial and lateral malleoli, the tibial spine, the tibial tuberosity and the lateral and medial plateaus were acquired on the tibia.
Using the anatomical landmarks, the navigation system was able to automatically identify the femoral mechanical axis, surgical trans-epicondylar axis (TEA), WSL, posterior condylar line for femur and tibial mechanical axis and the line connecting medial and lateral tibial plateau for tibia (Fig. 1).
For each patient, intra-operative kinematic acquisitions were collected before and after the definitive implant positioning. Pre-operative kinematic tests were specifically acquired after skin incision to allow the fixation of the tibial and femoral trackers, after medial parapatellar arthrotomy, before patella luxation and meniscal and anterior cruciate ligament (ACL) removal, while post-implant kinematic acquisitions were collected after the cementation of definitive prosthesis. Both the pre-operative and the post-operative acquisition were acquired with the tourniquet inflated, the joint capsule open and patella reduced. The kinematical data were acquired performing flexion-extension movements (full extension-full flexion-full extension), three times for each subject in two different conditions (Fig. 2): the passive motion (pROM), manually performed by the surgeon, maintaining the foot in neutral position (i.e. not introducing any additional stress/torque at foot level during the flexion-extension movement), and the active movement (aROM), directly performed by the patient.
Surgical technique
All the surgeries were performed under combined spinal and epidural anesthesia (CSE technique) which is a well-known technique typically used during labor. It offers the benefits of rapid onset of analgesia and at the same time allows lower-limb motor power [17]. Therefore, with the use of CSE anesthesia, the patients were able to perform active knee flexion-extension movement after skin incision and tracker positioning for navigation without experiencing pain. A midline skin incision was performed, and both femoral and tibial tracker were positioned in order to not interfere with surgical technique and prevent accidental mobilization. A standard medial parapatellar arthrotomy was performed and the patella was everted. Menisci and ACL were resected, and a tibial cut was made sparing PCL. After the cut of distal femur, the 4-in-1 guide was used to complete the femoral cut with the opportune size. The trial components were positioned and the flexion-extension gaps opportunely balanced, when needed. After the patellar cut and pulsed washing, definitive prothesis was implanted and the tourniquet finally released.
All patients were operated with the standard technique (medial parapatellar approach, adjusted mechanical alignment) and received a cemented Cruciate Retaining (CR) highly congruent Mobile Bearing (MB) TKA (Gemini, Waldemar LINK GmbH & Co. KG, Barkhausenweg 10, 22,339 Hamburg, Germany) with patella resurfacing.
Data analysis
The coordinate reference system on femur was defined as follows: the femoral mechanical axis as the proximal-distal (PD) axis, the anterior-posterior (AP) axis as the cross product between the PD-axis and the surgical TEA, and the cross product between AP-axis and PD-axis as the medial-lateral (ML) axis, thus achieving an anatomic orthogonal reference system.
The anatomic orthogonal reference system of the tibia was defined as: the PD-axis set as the tibial mechanical axis, the ML-axis as the cross product between the line connecting tibial spine and tibial tuberosity and the PD-axis, and the AP-axis as the cross-product between PD-axis and ML-axis.
Based on the acquired flexion-extension movements, the internal-external (IE) rotations were plotted against knee flexion. The AP translation was computed for both the medial and lateral epicondyles, evaluating their displacement projected in the transverse plane on the tibial reference system.
Statistical analysis
Starting from the analysis of literature [1, 2], a priori power analysis for a two-tailed paired Student’s t-test (alfa = 0.05, power = 0.8, mean difference of 3.0 ± 5.0° of rotations and 3.0 ± 5.0 mm of displacements) indicated a minimum sample size of 24 subjects.
For statistical comparison of the kinematic behavior, continuous data obtained from passive and active movements from 0° to 120°, both in pre- and post-operative conditions, were re-sampled each 5° of knee flexion using a smooth curve-fitting function that enabled direct comparison of patient.
IE rotations and AP translations values were then averaged on the three repetitions, at every re-sampled angle. The mean values obtained for each subject were then averaged for the whole cohort, thus obtaining one mean curve for the active condition and one for the passive one.
Both in pre- and post-implant conditions, internal-external (IE) rotations and anterior-posterior (AP) translations were estimated for pROM and aROM kinematic tests.
Differences between pre- and post- implantation and between active and passive motions, were statistically analysed using paired Student t-tests (p = 0.05). Statistical significance was set at 95% (p = 0.05). Analyse-it software (Analyse-it Software, Ltd., The Tannery 91 Kirkstall Road, Leeds, LS3 1HS, United Kingdom) was used to perform the reported statistical analysis.